I'm learning Haskell. I'm doing this for many reasons, but the most important is to expand my knowledge and form a more principled understanding of languages. So, I'm reading a paper. It's a very good paper, but the thing most striking about it is its age. 1977! My father was five. It occurred to me that I read a lot of old things. I grab interesting books from book sales, yard sales, thrift shops, and the "Transfer Station & Recycling Center" (aka the dump). Most of the time they're garbage, but I sometimes find true gems. It's how I discovered Plato's The Republic, among other things. The best part about reading older books is that they often strip away decades of intellectual cruft that can accumulate around a subject, be it jargon or just gross mutation of ideas. Especially reading some foundational papers, or books that were written when a subject was still young and exciting, a very good historical perspective can be developed. I often find a conceptual clarity that most new books simply don't give.
On the other hand, I've yet to find old code that was truly worth reading. Old code is frequently buggy, poorly written, or just plain unfinished. No code remains dormant for too long. The only exception I can think of is TeX.
Observation bias? Perhaps. It could be that that's the only reason these books have survived so long. Try reading some older books and papers. Let me know how it goes. Have some suggestions for great books and papers, old or others? I'd like to hear those too.